Posts made by Ben Warfield

Update on this: I increased the php memory limit in php.ini from 128M to 1024M and the login history is working now. The only reason I haven’t tried this before is because I’ve never seen a need to do this in other applications. My fog userTracking table currently has ~25,000 rows. I have other applications with millions of rows in a table that still work fine on 128M of memory.

I’m curious about storage nodes and how they are selected for image downloads. I have a main fog server with 2 storage nodes. When I image one machine at a time, how does fog select which node to download from?

Over the summer when we are imaging hundreds of machines, all 3 servers are busy most of the time. But now that summer imaging is over and I just do one or two machines at a time, I’ve noticed that images pull from node 1 most of the time, occasionally from node 2, and almost rarely will they download from the master. Just curious what the logic is behind this or if it’s random.

I’m on Ubuntu 12.04

Any updates on this? School is back in session and we would like to view history…

Hi, just wondering what the status is of this issue? I have the same problem but I see a different error in the apache log:

[Tue Aug 12 08:23:04 2014] [error] [client 10.162.3.26] PHP Fatal error: Allowed memory size of 134217728 bytes exhausted (tried to allocate 83 bytes) in /var/www/fog/lib/fog/Host.class.php on line 308

My /var/www/fog/management/reports folder only has the “Hosts and Users.php” file

Thanks!

What problem did you find? I seem to have the same issue, it’s weird because a lot of my login/logout times appear to be rounded up or down by half an hour.

I’m not sure if this is a bug or feature request…? Every once in a while I add a host manually instead of quick/full registration. When I drop down the list of images, it is sorted by image ID instead of image name. Is this by design? If I edit an existing host, the list is in the correct order.

Hey Mark,
The only QoS policies we have in place are for VoIP phones so I don’t think that should have any effect.

I’m actually doing my testing on the core. I noticed the issue offsite in another building and then came back to the office to test at the source. Yep - the server and clients are both on the core switch. This switch does the routing between vlans. As soon as the clients move out of the server vlan, multicast suffers.

I might have to look into the Iperf tests - it sounds like a neat tool. All other traffic seems okay though - I can copy files from PCs around the network and still get gigabit speeds.

Thanks for the thoughts! At this point the issue is only slowing me down a little bit - we’re just unicasting everything at the moment so it’s not a major problem, just something I’d like to try to fix.

Edit: forgot the IOS version, it is 12.2(17r)S4

We manage our switches directly through the CLI and I don’t have the CNA program to look at. Portfast is enabled on all the access ports and IGMP Snooping is enabled globally.

It is a Catalyst 6513

I saw this a few days ago on one of our machines and it ended up being because the new drive didn’t have a partition table. I used G-Parted to create an NTFS partition and did a quick format, then fog worked fine. Give that a try?

I’m having trouble getting multicast across vlans to work on our cisco network. I know this is more of a cisco question but thought I’d start here.

When I set up a multicast session to machines on the same vlan as the fog server, I get speeds above 3GB/Min. On any other vlan the speed drops to ~5MB/Min

I did search the forum and found this post: [url]http://fogproject.org/forum/threads/multicast-does-not-work-to-multiple-clients-only-to-single-client.863/#post-4316[/url]
I’ve tried changing the multicast address to the 239.x.x.x but it did not make a difference.

ip multicast-routing is turned on
the vlans have ip pim sparse-dense-mode enabled

Anyone have any suggestions?

Ahh okay, that makes sense. So the download button is basically just a shortcut. I’m usually using the download.

I would like to know the difference between the different deployment methods for a group. Is this documented anywhere? Specifically, what is the difference between the “Deploy” and “Download” buttons in the group listings? I understand what multicast is, but for the other 2, I don’t quite see a difference or reasoning behind one or the other.
Thanks!

Update: Setting the service to delayed start works perfectly. Thanks for the idea!

GFm - yep, I downloaded the service installer right from the new version of FOG

Junkhacker - hmm, good suggestion. I will try that and get back to you.

Has anyone else seen this issue or come across an answer? I’m still having to restart the fog service before it will change the hostname.

I’d like to echo what GFm has said, and provide another little detail.

I am also seeing the issue where the hostname changer does not work until the service is restarted. I will image a machine, it’ll come up and just sit there. When I look at the log, I’m seeing the exact same errors:

5/27/2014 4:19 PM FOG::HostnameChanger Attempting to connect to fog server…
5/27/2014 4:19 PM FOG::HostnameChanger Module is disabled on this host.
5/27/2014 4:19 PM FOG::HostnameChanger Host name was not found in the database.

However, I seem to have narrowed it down to a specific scenario: It only happens to me if the hostname has been changed in FOG. I’m testing on 3 machines and can reproduce this. Let’s say I have the machines named pc-1, pc-2, and pc-3. If I change pc-3’s name to be pc-10, I can re-image all 3 machines and the 3rd one will have the hostname changer problem. It requires me restarting the service to complete. I have verified that the hostnamechanger service is enabled.

Hope this helps…

Update: I got it to work.

The command to create the init.xz file is this:
xz -C crc32 -z -c init > init.xz

I found the answer right here in the forum: [url]http://fogproject.org/forum/threads/latest-fog-0-33b.6476/page-30[/url]
Thanks Tom!

Hi,
In fog version 0.32, I had modified init.gz to create a custom input option for auto registration. I would like to do the same for version 1.0.1 but am having trouble with the new init.xz file. I’ve edited FOGMountBootImage to reflect the new file location and to use xz instead of gzip and I attached what I’ve done below.

My problem is when I go to boot a machine, I get a message saying:
“Input was encoded with settings that are not supported by the XZ decoder”

I’m guessing I need to add more options to the command to recreate the compressed file, but I’m not sure what to add?

Thanks

[url=“/_imported_xf_attachments/0/789_FOGMountBootImage.txt?:”]FOGMountBootImage.txt[/url]